Smart and Automated Sewer Pipeline Defect Detection and Classification
Bibliographic record
Abstract
Currently, the condition of a sewer pipe is assessed by an inspector monitoring live video supplied from a remotely controlled closed-circuit television (CCTV) camera. As the inspector guides the video camera through the pipe, she/he will look for different types of defects/anomalies, including structural, operational, construction features, and miscellaneous defects. Based on the National Association for Sewer Service Companies (NASSCO) standard, there are 224 different defects/sub-defects which can occur within a given inspection. Given the significant number of defects/sub-defects, assigning defect codes and their corresponding severities is prone to subjectivity and hence may impact the overall accuracy of the inspection interpretations. Inaccurate interpretations could mislead decision makers while selecting the proper intervention actions to sustain critical sewers. In an effort to speed up the overall inspection process and enhance the interpretation accuracy, this research aims at utilizing artificial intelligence and computer vision tools to detect and classify defects in accordance with existing standards; this research is a continuation of AECOM X Google Hack-a-thon’s proof of concept application. The smart and automated tool relies on enormous data obtained from the City of Toronto, multiyear program to build a reliable database. The initial results of the prototype showed promising detection and classification capabilities of defects and sub-defects including circumferential crack (CC), longitudinal fracture (FL), encrustation attached deposits (DAE), tab break-in (TB), tab break-in intruding (TBI), and obstruction intruding (OBI). The average accuracy achieved for the six anomalies was 85% where the maximum and minimum accuracy levels were 94% and 75%. This tool, once completed, will elevate the sewer inspection process by speeding up the inspection validation, enhancing accuracy, and maintaining consistency, thereby assisting in making proper decisions when selecting the required intervention actions.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.004 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".